For engineering teams evaluating Viewrobe. How the try-on attaches to a product page, what the REST API does, and the security posture. If you want the shorter route through everything we publish, start at documentation. Covers the proprietary AI rendering pipeline, asynchronous, non-blocking deployment model, and security posture.
The rendering pipeline that transforms 2D product photography into hyper-realistic garment overlays, without 3D assets, manual retouching, or GPU requirements on the client.
The engine interprets the garment from a single 2D input: where it is structured and where it is loose, how it is constructed, how the material behaves. Fabric properties are inferred before the render pass begins, with no 3D asset to build per style.
A multi-stage pose detector locates the body in the shopper's photo and maps its geometry, so the garment can be placed correctly on a body that is turned, leaning or partly occluded rather than standing square to the camera.
Final render composites garment, skin-boundary blending, matched lighting, and ambient shadow in a single pass. Output supports PNG, WebP, and JPEG at resolutions up to 4K.
Asynchronous, non-blocking deployment. The client script attaches to your existing DOM structure without layout shifts or render-blocking requests.
A single asynchronous script tag attaches the VTON widget to any product page. The script detects your DOM structure, identifies product image containers, and injects the try-on trigger, zero configuration required.
Rendering happens off the host page. Asynchronous processing keeps the product page interactive while the try-on completes in the background, so the try-on never sits in the critical path of the page load.
Programmatic access for headless commerce stacks: submit product photography, request a render, and read back status. The route list is issued with API access at onboarding rather than published here.
Built for enterprise procurement. A short, defined retention window for user photos, encrypted transit, and infrastructure run on Google Cloud.
Fully GDPR and CCPA compliant. User photos are used only to generate the requested try-on, are permanently deleted within 24 hours, and are never used to train AI models. All data in transit uses TLS 1.3.
Built on Google Cloud, whose infrastructure is independently certified to SOC 2 Type II and ISO 27001. These are Google's certifications, not Viewrobe's. TLS 1.3 in transit and AES-256 at rest.
Hosted on Google Cloud. Support terms, including response times and any uptime commitment, are agreed per contract rather than published here. Viewrobe does not advertise an uptime figure or SLA.
Last updated: October 2026
Book a technical walkthrough with our integration engineering team.
Book demo